Using AI to Streamline the Data Collection Essential for Calculating Greenhouse Gas Emissions SMBC x Allganize Solution “Alli for Green”

SMBC has been offering the GHG emissions calculation tools Sustana and Persefoni since 2022, but when we use these tools to calculate our own emissions, first we have to collect the necessary data from power, gas, and other bills, meter readings, and other documents for the company as a whole and then input usage into these digital tools. As a result, even SMBC Group’s own calculations have required a lot of time and effort. We developed Alli for Green to simplify that part of the process.

Specifically, when you upload documents received in paper or PDF form to the cloud, Alli for Green automatically extracts preset items and displays them on the screen. The extracted items include, for example, the type of energy (power or gas), usage, and usage period. And that data can then be output as an Excel or CSV file, eliminating the need for manual data input.

Alli for Green uses Allganize’s natural language understanding AI to automatically extract the data needed for emissions calculations. Its major point of difference from other companies’ services is that it can accurately extract the necessary data even from documents with different formats. The old AI-OCR (optical character recognition using AI) method in popular use struggles with accuracy when items appear in different places in the document or where there is an inconsistency in notation, and the only way to improve that accuracy is to preset each individual item. Alli for Green, however, can extract data extremely accurately with just one click and with none of that complicated presetting.

That’s right. Another strength is our AI’s learning function. Just like humans, AI can make mistakes. The solution in such cases is to make the AI smarter, but that usually means gathering a huge amount of data and putting time and money into retraining the AI. Alli for Green, however, has an inbuilt retraining function enabling users to retrain their AI to be smarter with the push of a button. It’s designed so that accuracy automatically improves the more the AI is used. To the best of our knowledge, no other service offers this function.

For businesses that lease premises in a building, the various invoices created by the building owner in relation to energy usage, for example, often have major differences in their formats. Where AI-OCR has a lot of difficulty dealing with these discrepancies, Alli for Green extracts the data based on a contextual understanding, so it can read documents with a high degree of accuracy even if the information appears in an inconsistent notation and a different place in the document. We see this as a real game-changer.

A business of the size of SMBC handles documents with 1,000 different formats. Inputting them all manually is a huge task, but Alli for Green can do the whole job with the click of a button.

SMBC had already spent a couple of years looking into providing a solution that could resolve customers’ decarbonization issues. Decarbonization is an area that lends itself to digital streamlining in that it entails collecting massive amounts of data, building specialty logic, and then implementing calculations. SMBC has been pursuing service development in this area, as well as working with greentech startups.

In the course of tackling GHG emissions calculation as the first step toward decarbonization, SMBC heard from customers that data collection at the stage prior to using calculation tools was imposing a real burden, and we felt that SMBC too needed to streamline that process. We have branches all over the country, each of which receives reports and invoices on energy use. In many cases, our branches lease their premises, and they receive invoices in a whole range of formats. We decided to work with Allganize to develop a digital service that would streamline data collection and digitalization with an eye to extending the service to customers facing the same issue.

Allganize is a startup that SMBC Group’s Silicon Valley Digital Innovation Lab got to know in Silicon Valley. Seeing major possibilities in the company’s advanced natural language understanding technologies, we decided to become an investor. When we subsequently introduced Allganize products—its AI chatbot, for example—we were very impressed with the superiority of the company’s natural language processing technologies, its development speed, and its agility, so we approached Allganize with the idea of applying its advanced technologies to the decarbonization field. Allganize told us that they had a generic AI technology that could understand context and extract specific data regardless of the format. That sounded promising as a means of resolving the data collection problem, which is how we ended up working together on this project.

To use Persefoni to calculate its GHG emissions, SMBC has to input data extracted by Alli for Green into the Persefoni software, but that input process was initially so time-consuming that the workload was not going to be substantially reduced. After a process of trial and error, we added a function mapping that data to Persefoni’s bulk upload format so that the format can be created with one click and the data dropped into the calculation tool.

We approached development with a priority on user-friendliness, liaising closely with the SMBC departments that would actually be using Alli for Green. In terms of provision of the product to customers too, while the service itself is an Allganize product, it would be SMBC that proposed the service to its customers, so we checked the quality meticulously before the service release.

As far as I know, no other company offers the same service. Everyone is using GHG emissions calculation tools, but many companies have problems at the data collection stage prior to calculation. Some calculation tools can read figures from invoices and other documents, but Alli for Green has an overwhelming advantage over these in terms of accuracy. We’ve received a lot of enquiries about the service, particularly from Japan and the United States.

Both Sustana and Persefoni have seen a growing number of enquiries and signups over the last year, and the government has recommended that listed companies in Japan include their Scope 1 and 2 emissions* in their Annual Securities Reports as of FY2023, so there seems to be a growing awareness of corporate decarbonization management.

Yes. With so many big companies now using calculation tools, their next focus will most likely be on streamlining the prior stage of data collection, so I think Alli for Green has great potential.

SMBC Group aims to provide not just financial services but also high-quality solutions that address customers’ needs and social issues, but handling all our development inhouse is impossible, particularly in terms of achieving the necessary development speed. That’s where it would be great to continue combining Allganize’s speed and technologies with SMBC Group’s infrastructure to create new value.

SMBC Group actually moves very quickly for a big company and it also has a deep understanding of agile development, which is what enabled this project to proceed so smoothly. It was thanks to SMBC approaching us that we realized that our technology also has a role in the decarbonization field. I might add that there is huge social value in SMBC Group as one of Japan’s mega-bank groups acting as a decarbonization frontrunner. I would like to extend business use cases like this, where our natural language understanding AI technology is utilized to enable people to focus on high-value tasks that only people can perform, beyond Japan out to the United States and South Korea, where we also have operations, and indeed worldwide.

I would like SMBC as a megabank group to continue to pursue decarbonization management while also supporting other companies’ decarbonization management through the provision of Alli for Green and other services. I also look forward to the Group working with other financial institutions and startups to become a global solutions provider that can offer social value in fields beyond decarbonization as well.